Significance Although vitamin A is an essential nutrient for normal cellular function, recent epidemiological studies have suggested, but not proven, that ingestion of excess vitamin A supplements during early pregnancy may cause birth defects. Objectives The present study was carried out to determine safe levels of vitamin A during pregnancy in the cynomolgus monkey which can be used for risk assessment. The cynomolgus monkey, a well established model for normal and abnormal human development, mimics the human malformation syndrome associated with the vitamin A analog, isotretinoin (Accutane). Results Oral exposure to excessive doses of vitamin A (retinyl palmitate, AROVIT) administered at 7,500 to 80,000 IU/kg on gestational days 16-27 resulted in dose-related increases in abortions and malformations which affected typical retinoid target tissues, including the craniofacial region, heart, and thymus. The no observed adverse effect level (NOAEL) and lowest observed adverse effect level (LOAEL) were 7,500 and 20,000 IU/kg, respectively. Based on the similarities between monkeys and humans in response to isotretinoin, the monkey data can be used to extrapolate a hypothetical safe dose of vitamin A in humans. Thus, considering weight differences (4 kg monkey and 50 kg woman), the monkey NOAEL, which is equivalent to 30,000 IU/day can be extrapolated to a theoretical human NOAEL of ~300,000 IU/day vitamin A. Since this level is ~100x the RDA (2,670 IU) for vitamin A, these results indicate that the current recommendation for vitamin A use during pregnancy provides a large margin of safety. Future Directions Use of the monkey retinoid model to study the molecular mechanisms of craniofacial dysmorphogenesis induced by excess vitamin A. KEYWORDS vitamin A, birth defects, retinoids, risk assessment